Is Yuka Accurate? Half the Score Is Proven, Half Isn't

Yuka’s 100-point score is not one claim. It is two claims stapled together, and only one of them has been tested against real health outcomes in hundreds of thousands of people. The nutrition math behind the score sits on a body of cohort evidence that keeps replicating across countries. The additive math sits on a mix of regulatory opinions and individual studies that the agencies writing them openly call uncertain. Reading a red “high risk” additive flag with the same confidence as the green nutrition grade is the actual myth here, not whether Yuka is “accurate” in some single sense.

How the 100-point score is built

Yuka is transparent about its formula, which makes it easy to pull apart.

• Nutritional quality: 60 percent. Yuka’s own adaptation of a nutrient profiling model in the same family as the Nutri-Score front-of-pack label.
• Additive risk: 30 percent. Each additive gets sorted into a risk tier, from no risk to high risk, based on regulatory opinions and published studies that Yuka selects.
• Organic status: 10 percent. A flat bonus for organic certification.

Those three numbers get blended into one color and one score. A shopper sees a single verdict, not three separate ones with three separate confidence levels.

The cohort science behind 60 percent of the score

The nutrition component is not a guess. The model it is built on, the Food Standards Agency nutrient profiling system that also underlies Nutri-Score, was tested against mortality outcomes in the EPIC cohort of more than 500,000 adults across 10 European countries.^[1]

7% higher all-cause mortality risk for the worst-scoring fifth of diets versus the best fifth, EPIC cohort PMID 32938660

The number that matters: people whose diets scored in the worst fifth of the nutrient profile model had a 7 percent higher all-cause mortality risk than those in the best fifth, tracked over a median of 17 years.^[1] That is not a rounding error. It is also not a dramatic number, and it should not be sold as one.

What a cohort study can and cannot prove

Here is the part the app’s clean red-to-green gradient tends to flatten.

The EPIC findings come from a population-based cohort study that measured participants’ habitual diets with dietary questionnaires. Nobody was randomly assigned to eat a particular way.^[1] That is an observational design. It supports an observational claim: a pattern that shows up across a population, not proof of what caused any one person’s outcome.

What the cohort proves: that people whose overall diet pattern scores worse on the nutrient profile model die at a somewhat higher rate. What it does not prove: that swapping one Yuka-flagged product for a green one changes any single shopper’s personal risk. A population association is not a personal guarantee.

That distinction matters because it is exactly the kind of nuance a single number cannot carry. A credit score works the same way: it predicts default rates across a population of borrowers with real accuracy, but it cannot tell one specific borrower what will happen to them.

Where the other 30 percent leans on thinner ground

The additive tier is a different kind of evidence entirely. It is assembled from regulatory safety opinions and individual published studies, and those do not carry the same weight or replication as a multi-country cohort outcome study.

Titanium dioxide, additive E171, is the clearest example. In its 2021 re-evaluation, the EU’s food safety agency EFSA concluded that a concern for genotoxicity of titanium dioxide particles could not be ruled out, and that E171 could no longer be considered safe as a food additive, a finding that led to its withdrawal from the EU’s approved additive list.^[2]

That sounds decisive. It is worth reading the actual basis for it. Titanium dioxide was removed from the European market while direct human evidence remained limited. As of 2026, one of the first human dietary intervention studies on E171 involved just 31 adults over two weeks.^[3] The genotoxicity concern that drove the ban was built mostly on cell and animal data, not on evidence of harm measured directly in people.

The two halves of a Yuka score, compared

Same 100-point number, different evidence underneath.

None of this means additive titanium dioxide is safe, or that EFSA got it wrong. It means a “high risk” flag on the additive side of the score is a different kind of statement than a low grade on the nutrition side, even though the app hands the reader one color for both.

What no one has measured

No published study has tested whether a shopper’s individual Yuka score predicts their personal health outcomes. That gap is not a scandal. It is just a gap, and it is worth naming plainly.

The cohort evidence validates the nutrient profiling logic behind the nutrition component in aggregate. It does not validate the finished 100-point number that any single app displays on any single product, because no one has run that study.

What to do with the score

The nutrition component is the part of the score with the strongest population-level evidence behind it, and it is the most defensible reason to prefer one product over another when two options are otherwise similar.

A high-risk additive flag is worth reading differently.

• Treat it as a prompt. Look up the specific regulatory opinion behind the flag before treating it as a settled verdict.
• Ask what kind of evidence backs it. A human cohort finding and a rat study measuring DNA strand breaks are not interchangeable, even when an app shows them in the same red.
• Remember the score is a population tool. It was built and tested at the level of large groups. Individual shoppers were never the unit of measurement.

For a fuller side-by-side of how scanner apps like this one differ in method, see Best food scanner apps 2026: EatMatrix vs Yuka, on the evidence. It is also why the EatMatrix app links every additive verdict straight back to the regulator’s own assessment, instead of collapsing it into one unexplained color.

Bottom line: Yuka’s nutrition grade rests on real, repeated cohort science. Its additive grade rests on a patchwork of regulatory calls and preclinical studies that even the regulators describe as uncertain. Asking “is Yuka accurate” without separating those two halves is asking the wrong question. The right one is which half you are looking at, and how much certainty that half has earned.

This article summarises publicly available regulatory assessments and peer-reviewed research about food ingredients. It is general information and is not intended to diagnose, treat, cure, or prevent any disease, and it is not personalised dietary or medical advice. Regulatory positions change; the EatMatrix app reflects the current assessments in our database. If you have an allergy, a medical condition, or are pregnant, talk to a qualified clinician or dietitian before changing what you eat.